Results
Disperse workflow results
Data description
The data were obtained from Quijote simulations: Big Sobol Sequence. The following results are based on the snapshot at /Snapshots/BSQ/0/snap_010.hdf5.
Replicate disperse tutorial on point sample: 3D walls and filaments
See the disperse version here. All visualization is done with Paraview. For each analysis the Python commands with arguments are listed below.
Thinned sample (of the 134M points keep 2M points)
python scripts/analyze_snapshot.py \
--input data/snap_010.hdf5 \
--output-dir outputs/snap_010_thinned \
--target-count 2_000_000 \
--delaunay-btype periodic \
--export-delaunay \
--mse-nsig 3.0 \
--dump-manifolds JE1a \
--dump-arcs U \
--netconv-smooth 20 \
--skelconv-smooth 20Full cube. Smoothed 20x, walls and filaments. NDnet as points.
Clip at x = 50 and zoom in. Smoothed 20x, walls and filaments. NDnet as points.
Cropped sample (0.5 x 0.5 x 0.1 Gpc box) at full resolution (smoothed at the end)
python scripts/analyze_snapshot.py \
--input data/snap_010.hdf5 \
--output-dir outputs/snap_010_cropped \
--crop-box 0 0 0 500000 500000 100000 \
--stride 1 \
--delaunay-btype periodic \
--export-delaunay \
--mse-nsig 3.0 \
--dump-manifolds JE1a \
--dump-arcs U \
--netconv-smooth 20 \
--skelconv-smooth 20Full cube (thinned) and crop (at full resolution). Smoothed 20x, walls only. The full cube is based on a thinned sample of 2M points. The crop is a subbox of the full 134M points.
Full cube and crop zoom front. Smoothed 20x, walls only.
Full cube and crop zoom side. Smoothed 20x, walls only. Notice the difference in resolution compared to the thinned sample.
Full crop. Smoothed 20x, walls and filaments. NDnet as points.
Zoom crop. Smoothed 20x, walls and filaments. NDnet as points.
Replicate disperse tutorial on point sample: 2D voids and filaments
See the disperse version here.
Thinned sample (of the 134M points keep 2M points)
python scripts/analyze_snapshot_2d.py \
--input data/snap_010.hdf5 \
--output-dir outputs/snap_010_thinned_2D \
--projection-axis z --projection-range 0 10000 \
--target-count 2_000_000 \
--delaunay-btype periodic \
--export-delaunay \
--mse-nsig 3.0 \
--mse-vertex-as-minima \
--dump-manifolds J0a \
--dump-arcs U \
--netconv-smooth 20 \
--skelconv-smooth 20Full plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet as surface.
Full plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet and filaments as surface.
Zoomed plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet and filaments as surface.
Full plane, z-projection 0-10000. Smoothed 20x. Source index as gaussian points and filaments as surface.
Zoom plane, z-projection 0-10000. Smoothed 20x. Source index as gaussian points and filaments as surface.
Cropped sample (0.5 x 0.5 x 0.1 Gpc box) at full resolution (smoothed at the end)
python scripts/analyze_snapshot_2d.py \
--input data/snap_010.hdf5 \
--output-dir outputs/snap_010_cropped_2D \
--projection-axis z --projection-range 0 10000 \
--crop-box 0 0 0 500000 500000 100000 \
--stride 1 \
--delaunay-btype periodic \
--export-delaunay \
--mse-nsig 3.0 \
--mse-vertex-as-minima \
--dump-manifolds J0a \
--dump-arcs U \
--netconv-smooth 20 \
--skelconv-smooth 20Crop plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet as surface.
Crop plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet and filaments as surface.
Zoomed crop plane, z-projection 0-10000. Smoothed 20x. Delaunay NDnet and filaments as surface.
Crop plane, z-projection 0-10000. Smoothed 20x. Source index as gaussian points and filaments as surface.
Zoom crop plane, z-projection 0-10000. Smoothed 20x. Source index as gaussian points and filaments as surface.